/**
 * Options for speech synthesis
 */
export interface SpeakOptions {
    /**
     * Speech rate
     */
    rate?: "x-slow" | "slow" | "medium" | "fast" | "x-fast";
    /**
     * Speech pitch
     */
    pitch?: "x-low" | "low" | "medium" | "high" | "x-high";
    /**
     * Speech volume (0-100)
     */
    volume?: number;
    /**
     * Whether to convert input from SpeechMarkdown to SSML before processing
     */
    useSpeechMarkdown?: boolean;
    /**
     * Whether to use word boundary information for streaming synthesis
     */
    useWordBoundary?: boolean;
    /**
     * Path to save the audio output
     */
    outputPath?: string;
    /**
     * Voice ID to use for synthesis
     */
    voice?: string;
    /**
     * Audio format to use for synthesis
     */
    format?: "mp3" | "wav" | "ogg" | "opus" | "aac" | "flac" | "pcm";
    /**
     * Raw SSML to pass directly to the provider, bypassing Speech Markdown conversion
     * This allows users to use provider-specific SSML features that may not be supported
     * by the Speech Markdown library or the wrapper's SSML processing
     */
    rawSSML?: boolean;
}
/**
 * Input source for speak methods - can be text, file, bytes, or stream
 * Only one of these should be provided
 */
export interface SpeakInput {
    /**
     * Text or SSML to synthesize and speak
     */
    text?: string;
    /**
     * Path to an audio file to play
     */
    filename?: string;
    /**
     * Raw audio bytes to play
     */
    audioBytes?: Uint8Array;
    /**
     * Audio stream to play
     */
    audioStream?: ReadableStream<Uint8Array>;
}
/**
 * Unified voice format across all providers
 */
export type UnifiedVoice = {
    /**
     * Unique identifier for the voice
     */
    id: string;
    /**
     * Display name of the voice
     */
    name: string;
    /**
     * Gender of the voice
     */
    gender?: "Male" | "Female" | "Unknown";
    /**
     * TTS provider
     */
    provider: "azure" | "cerevoice" | "google" | "ibm" | "elevenlabs" | "polly" | "witai" | "playht" | "openai" | "upliftai" | "modelslab" | "sherpa" | "sherpaonnx" | "sherpaonnx-wasm" | "espeak-ng" | "sapi" | "cartesia" | "deepgram" | "hume" | "mistral" | "murf" | "resemble" | "unrealspeech" | "xai" | "fishaudio" | "gemini";
    /**
     * Language codes supported by this voice
     */
    languageCodes: {
        /**
         * BCP-47 language code (e.g., 'en-US')
         */
        bcp47: string;
        /**
         * ISO 639-3 language code (e.g., 'eng')
         */
        iso639_3: string;
        /**
         * Human-readable language name (e.g., 'English (United States)')
         */
        display: string;
    }[];
    /**
     * Provider-specific metadata
     */
    metadata?: {
        /**
         * AWS Polly supported engines (standard, neural, long-form, generative)
         */
        supportedEngines?: string[];
        /**
         * Additional language codes for bilingual voices
         */
        additionalLanguageCodes?: string[];
        /**
         * Any other provider-specific data
         */
        [key: string]: any;
    };
};
/**
 * Credentials for TTS providers
 * @remarks This is intentionally kept as a generic object type to allow different credential formats for different providers
 */
export type TTSCredentials = Record<string, unknown>;
/**
 * Event types for TTS callbacks
 */
export type TTSEventType = "start" | "end" | "boundary";
/**
 * Callback function for word boundary events
 */
export type WordBoundaryCallback = (word: string, startTime: number, endTime: number) => void;
/**
 * Simple callback function with no parameters
 */
export type SimpleCallback = () => void;
/**
 * Property type for TTS properties
 */
export type ModelFeature = "streaming" | "audio-tags" | "inline-voice-cloning" | "open-source" | "word-boundary-events" | "character-boundary-events" | "ssml";
export declare const FEATURES: Record<string, ModelFeature>;
export interface ModelInfo {
    id: string;
    features: ModelFeature[];
}
export type PropertyType = string | number | boolean;
/**
 * Word Boundary interface for speech marks
 */
export type WordBoundary = {
    /**
     * The actual word text
     */
    text: string;
    /**
     * Time offset from the beginning of the audio stream, in milliseconds.
     */
    offset: number;
    /**
     * Duration of the word being spoken, in milliseconds.
     */
    duration: number;
};
/**
 * Detailed credentials check result
 */
export type CredentialsCheckResult = {
    /**
     * Whether the credentials are valid
     */
    success: boolean;
    /**
     * Error message if credentials are invalid
     */
    error?: string;
    /**
     * Number of voices available if credentials are valid
     */
    voiceCount?: number;
};